A Kinetic Study of Nickel Complexation in Model Systems by

Feb 6, 2001 - Adsorptive cathodic stripping voltammetry (AdCSV) in conjunction with the competing ligand-exchange method (CLEM) was investigated as a ...
1 downloads 6 Views 82KB Size
Download PDF
Environ. Sci. Technol. 2001, 35, 1084-1089

A Kinetic Study of Nickel Complexation in Model Systems by Adsorptive Cathodic Stripping Voltammetry VALBONA CELO, JOHN MURIMBOH, MOHAMED S. A. SALAM, AND CHUNI L. CHAKRABARTI* Ottawa-Carleton Chemistry Institute, Department of Chemistry, Carleton University, 1125 Colonel By Drive, Ottawa, Ontario K1S 5B6, Canada

Adsorptive cathodic stripping voltammetry (AdCSV) in conjunction with the competing ligand-exchange method (CLEM) was investigated as a tool for measuring dissociation rate coefficients of nickel complexes in model systems. Dimethylglyoxime (DMG) was used as the competing ligand. Citric acid (CA) and a well-characterized fulvic acid (FA) were used as model ligands. The rate coefficients were calculated, and the consistency of equilibrium and kinetic data was discussed. The contributions of the disjunctive pathway (proceeding by the dissociation of the initial complex) and the adjunctive pathway (proceeding by the formation of an intermediate complex as a result of direct attack of the competing ligand on the initial complex) on the overall reactions were investigated. The reactions of NiCA or Ni-FA complexes with DMG were demonstrated to proceed by both disjunctive and adjunctive pathways. The predominant pathway for the overall reaction depends on the nickel-to-initial ligand and the DMG-to-initial ligand ratios. The reactions follow predominantly the disjunctive pathway for [DMG] g 3 mM and Ni-to-dissolved organic carbon (DOC) ratios greater than 10 nM Ni 2+/g of DOC. Since free nickel ion in freshwaters is reported to be toxic, its rate and pathway of formation are of environmental concern.

Introduction Interactions between transition metal ions and a wide and complex variety of organic ligands, inorganic anions, reducible and oxidizable chemical species, surfaces, and organisms in natural water systems play a significant role in the hydrogeochemical and biogeochemical cycle of these metals in the aquatic environment and are of great importance in understanding transport, toxicity, and bioavailability of metal species in natural waters. Both metal speciation and bioavailability are functions of the tendency of the metal ion to react, as quantified by the free metal ion activity, with organic matter under pseudoequilibrium conditions (1). Among the physical and chemical factors affecting the bioavailability of trace metals in the aquatic environment, complexation by natural (NOM) organic matter is important. * Corresponding author e-mail: [email protected]; telephone: (613)520-2600, Ext. 3839; fax: (613)520-3749/3830. 1084

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001

The main components of NOM in the aquatic environment are humic substances. Buffle (2) has reported that soil fulvic acids (FAs) comprise up to 75% of the NOM content in rivers and streams. Humic and fulvic substances are highly complex mixtures of polyelectrolytes with a wide range of functional moieties and molecular weight distributions (3). Binding of heavy metals by humic and fulvic acids strongly affects their availability to living organisms (4). Living systems form examples par excellence of ordered dynamic systems far from equilibrium (5). Additionally, the natural processes that determine the physical and chemical properties of aquatic system are not in equilibrium. Disequilibria between metals and complexing agents in natural waters may result from changes that arise from natural as well as anthropogenic inputs and from hydrological, geochemical, and biological processes (6). For such nonequilibrium systems, kinetic speciation would give a more realistic description of bioavailability of trace metals than metal speciation based on local equilibrium approximations. For nickel complexes in freshwaters, Cabaniss (7) has shown that relatively inert systems such as Ni-FA complexes require both kinetic and equilibrium models to predict metal partitioning properly. Adsorptive cathodic stripping voltammetry (AdCSV) with dimethylglyoxime (DMG) has been used for nickel determination (8-10) and equilibrium speciation (11, 12) in natural waters. Recent voltammetric studies indicate that under marine conditions, more than 95% of total nickel is organically complexed (11). Langford et al. have used the competing ligand-exchange method (CLEM) for kinetic speciation of nickel with the chromophore 4-(2-pyridylazo)resorcinol (PAR) as a competing ligand and spectrophotometric detection of the NiPAR complex (13). Chakrabarti et al. have studied the kinetic speciation of Ni-FA complexes using CLEM with Chelex-100 as the competing ligand and graphite furnace atomic absorption spectrometry (GFAAS) or inductivelycoupled plasma-mass spectrometry (ICP-MS) to measure the dissociation kinetics (14-17). However, there are only two publications (both from this laboratory) in which the AdCSV method has been used to determine the kinetic speciation of nickel in natural waters (18, 19). In all these studies, the authors have assumed that the ligand-exchange reactions of Ni-FA complexes involve the dissociation of the initial complexes as the rate-determining step and have determined the rate coefficients for dissociation of these complexes. The complexity and heterogeneity of naturally occurring organic complexants, such as humic substances, in freshwaters present problems in interpretation of kinetic results. It is therefore necessary to study metal complexation properties in model systems so that the results can be extrapolated from the laboratory to natural water conditions. In this study, model systems containing citric acid (CA) and a well-characterized FA solution as model ligands have been used. CA has been used earlier as a model ligand to simulate the discrete site models of binding by humic substances (20). The FA concentration in model systems is intended to simulate the DOC concentration of freshwaters. In this work, we have investigated rates of ligand-exchange reactions for nickel complexes in model systems using CLEM with DMG as the competing ligand and AdCSV to measure the dissociation kinetics. Two reaction pathways for the overall ligand-exchange reactions may be considered: the disjunctive pathway (dissociation of the metal complex) and the adjunctive pathway (direct attack by the incoming ligand to form an intermediate complex) (5, 6, 21-23). Herring and 10.1021/es001203q CCC: $20.00

 2001 American Chemical Society Published on Web 02/06/2001

Morel (21) introduced this new terminology to denote this as a stoichiometric mechanism (a sequence of elementary steps) as opposed to an intimate mechanism that deals with the activation of the rate-determining step. If the disjunctive pathway is followed, the measured rate coefficients can be related to a fundamental process in the natural environments dissociation of the nickel complex. However, the observed rate coefficients for the adjunctive pathway depend on the nature of an arbitrarily chosen probe ligand, and consequently, little information can be obtained about the processes in the natural environment. The theoretical and experimental conditions in which one of these pathways predominates have been analyzed.

If the overall reaction follows the disjunctive pathway, then disj

k1

} Ni2+ + L NiL {\ disj k-1

disj

k2

Ni2+ + 2DMG 98 Ni(DMG)2 Assuming a steady-state approximation for [Ni2+] (see Supporting Information), one can obtain:

kdisjunctive )

Theory Ligand-Exchange Reactions. The overall ligand-exchange reaction of the initial nickel complex, NiL, with DMG as the competing ligand can be described by koverall

NiL + 2DMG 98 Ni(DMG)2(sol) + L

(1)

The electrochemical measurements are based on adsorption of the Ni(DMG)2 complex on the mercury electrode (11, 12) and reduction of the adsorbed complex, which according to Ma et al. (9) is a stepwise reduction involving a total of 10 electrons:

Ni(DMG)2(sol) f Ni(DMG)2(adsorbed) -

0

Ni(DMG)2(adsorbed) + 10e f Ni + 2DMG

(2) (3)

Pihlar et al. (8) have reported that the reaction of Ni(DMG)2 reduction on a hanging mercury drop electrode is first-order with respect to the concentration of Ni2+. The peak current as measured by AdCSV is proportional to the concentration of Ni(DMG)2 adsorbed on the mercury electrode surface (9, 24). Provided that both adsorption of the Ni(DMG)2 complex from the aqueous solution on the mercury electrode surface (eq 2) (25) and reduction of the adsorbed complex (eq 3) are fast, the increasing rate of the measured peak current corresponds to the rate of formation of Ni(DMG)2(sol); hence, the rate coefficient calculated for this process is koverall of eq 1. The rate law for the overall reaction shown as eq 1 is given by

d[Ni(DMG)2] ) koverall[NiL][DMG]2 dt

where kobserved ) koverall[DMG]2. It can be shown that

ipt ) imax[1 - exp(-kobservedt)]

where [L] is the concentration of the initial ligand and k is the rate coefficient for the disjunctive pathway. In this case, kobserved will strongly depend on the relative concentrations of the initial ligand, L, and the competing ligand, DMG. If the reaction follows the adjunctive pathway, then

disjunctive

adj

k1

NiL + 2DMG {\ } (DMG)2 NiL adj k-1

kadj 2

(DMG)2NiL 98 Ni(DMG)2 + L In this case, the rate coefficient for eq 1 is given by eq 8 (see Supporting Information), and koverall is independent of the concentrations of both the initial ligand, L, and the competing ligand, DMG:

kadjunctive )

adj kadj 1 k2

(6)

(8)

adj k-1 + kadj 2

In general, the reaction given by eq 1 may follow both the adjunctive and the disjunctive pathways simultaneously, although one pathway may predominate under certain experimental conditions:

kobserved ) kadjunctive + kdisjunctive

(9)

If the concentrations of the competing ligand, DMG, and 2 the initial ligand, L, are such that the condition kdisj 2 [DMG] disj , k-1 [L] is satisfied, then eqs 7 and 9 can be rewritten as

kdisjunctive )

(5)

(7)

disj 2 kdisj 1 [L] + k2 [DMG]

(4)

Equation 4 expresses the rate law for the overall reaction, and no inference can therefore be drawn from eq 4 about the molecularity of elementary reactions. In the kinetic measurements, the condition that [DMG] . [Ni2+] is maintained. Since [NiL] < [Ni2+], [DMG] . [NiL], and eq 1 is pseudo-first-order, the integrated form of the rate law is

[Ni(DMG)2]t ) [NiL]t)0[1 - exp(-kobservedt)]

disj kdisj 1 k2

disj kdisj 1 k2

(10)

disj k-1 [L]

kobserved ) kadj +

disj kdisj 1 k2

(11)

disj k-1 [L]

In this case, a plot of kobserved versus 1/[L] will be a straight line with intercept equal to kadjunctive and the slope equal to disj disj kdisj 1 k2 /k-1 . The key hypothesis to be tested is that the observed rate coefficient of the overall reaction, kobserved, is equal to the dissociation rate coefficient of the initial complex NiL (called kdisj 1 ) only if

(i) the disjunctive pathway predominates where ipt is the peak current at time t and imax is the maximum current in the absence of complexants. The observed rate coefficient, kobserved, can be calculated by fitting the kinetic data to the exponential rise model in eq 6.

disj 2 (ii) kdisj 2 [DMG] . k-1 [L]

(12)

If the above two conditions are satisfied, then VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1085

d[Ni(DMG)2] d[NiL] )) kdisj 1 [NiL] dt dt and hence

kobserved ) kdisj 1 The disjunctive and adjunctive pathways can be distinguished experimentally by the dependence of the observed rate coefficient on [L]. If the reaction follows the adjunctive pathway, the overall rate coefficient will not depend on [L]. If the disjunctive pathway predominates, the overall rate coefficient will strongly depend on [L].

Experimental Section Apparatus. Voltammetric measurements were carried out using the Bioanalytical Systems (BAS) 100B/W electrochemical analyzer, with controlled-growth mercury electrode (CGME) stand (BAS), controlled by a Pentium 166MMX personal computer (Gateway 2000). A medium size (size 6) hanging mercury drop electrode was employed as the working electrode. The reference electrode was a Ag/AgCl electrode (BAS), filled with 3 M NaCl, and the counter electrode was a Pt wire (BAS). Analysis of voltammetric peaks was done using the BAS 100W windows control software v2.1 (BAS). The deposition potential was -700 mV, and the adsorption time was 5 s. Square wave anodic stripping voltammetry was used with a frequency of 100 Hz, a scan rate of 5 mV/s, and a pulse height 20 mV. All the measurements were performed under the nitrogen gas flow condition. Materials. A stock solution of 1000 µg/mL Ni(II) was prepared by dissolving pure nickel powder (SPEX, 99.999%) in ultrapure nitric acid (Baker Inc., ULTREX II), diluting with ultrapure water containing 1% (v/v) nitric acid. A 0.1 M DMG stock solution was prepared by dissolving an appropriate amount of solid DMG (Fisher Scientific, certified) in absolute ethanol (Spectro grade). The HEPES (BDH, ACS grade) and NaOH (BDH, ACS grade) reagents were used to prepare the aqueous stock solutions of 2 M HEPES and 2 M NaOH, respectively. The BHD analytical grade reagents were used to prepare the stock solution of CA. All solutions were prepared in ultrapure water of 18.2 MΩ resistivity obtained directly from a Milli-Q UF Plus water purification system (Millipore), fitted with a column to remove organic impurities in ultrapure water. The ionic strength was kept constant throughout this work at 0.02 M using NaOH. A well-characterized FA was collected from Armadale, Prince Edward Island, Canada, and was extracted from a Bh horizon by Dr. D. S. Gamble, Agriculture Canada, Ottawa (26, 27). The total number of phenolic and carboxylic groups determined by potentiometric titration is 3.9 and 7.71 mmol/g of FA, respectively (26-28). The bidentate complexing capacity is therefore approximately 5.4 mmol/g of FA. The Armadale FA stock solution (1.0219 g/L) was prepared by dissolving an appropriate amount of the freeze-dried FA in ultrapure water. The stock solution was stored in the dark at 4 °C until use. Kinetic Experiments. The test solutions for kinetic measurements were prepared by mixing appropriate amounts of nickel and the initial ligand, L (CA or FA) solutions. The mixtures were left overnight for equilibration at room temperature (22 °C). The ionic strength and the pH were held constant for all experiments by adding 0.200 mL of 2 M NaOH and 0.175 mL of HEPES 2 M solutions to 20 mL of the sample to fix the pH at 8.3 and the ionic strength at 0.02 M for the kinetic experiments. The kinetic experiments were started immediately upon addition of DMG to the test solution. The peak current was recorded every 25 s, and the 1086

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001

FIGURE 1. ip/imax ratio of nickel citrate solutions containing p various concentrations of citric acid as a function of time: [, no CA; b, [CA] ) 16.7; 0, [CA] ) 27.3; 9, [CA] ) 36.4; 2, [CA] ) 45.5 µM. recording was continued until no further change in the signal was observed and the current achieved a maximum value imax p . The CA model solutions contained 400 nM Ni2+ and 0.25500 µM CA. The kinetic measurements were performed using the DMG concentrations ranging from 0.5 to 3 mM. Model solutions of the FA were prepared to contain 400 nM Ni2+ and up to 200 mg/L FA. Data Analysis. The overall reaction was pseudo-firstorder, and the rate coefficients were calculated by fitting the experimental data to a one-component exponential rise model by nonlinear regression analysis using the MarquardtLevenberg algorithm. The calculations were performed using SigmaPlot 5 computer program (SPSS Science).

Results and Discussion Citric Acid Model Solutions. Figure 1 shows the kinetic curves, i.e., ip/imax (%) as a function of time for the test p solutions containing varying concentrations of CA (up to 50 µM) and 1 mM DMG. For the complexes of nickel with CA, the experimental conditions were such that only one complex, Ni-citrate, was formed (according to MINEQL calculations), and hence, the kinetic data were fitted to the one-component exponential rise model. The kinetic measurements were performed as described earlier, and the kobserved values and their range are presented in Table 1. Figure 2 shows the general trend and makes clear that the rate coefficient for the overall reaction strongly depended on the concentration of DMG and CA for [DMG] < 2 mM. With increasing concentrations of DMG, the measured rate coefficient became constant, regardless of the concentration of DMG or CA. This observation can be interpreted in two ways: either the adjunctive pathway disj 2 predominated (eq 8) or kdisj 2 [DMG] . k-1 [L], i.e., the condition shown by eq 12, was satisfied and the disjunctive pathway predominated. Further kinetic measurements were performed using experimental conditions such that the rate coefficient depended on the concentrations of CA or DMG so that we could determine the experimental conditions under which one of these two pathways predominated and to calculate the contribution of kdisjunctive and kadjunctive to the kobserved. The relationship expressed by eq 11 for the model systems of CA and DMG is presented in Figure 3. The nonzero intercept accounts for some contribution of the adjunctive pathway to the stiochiometric mechanism of the overall

TABLE 1. koverall (s-1) Calculated by Fitting Kinetic Data to One-Component Exponential Rise Model Using Nonlinear Regression Analysis, Applied for Ni-Citric Acid Model Systems koverall (s-1) citric acid (µM)

0.5 mM DMG

1.0 mM DMG

1.5 mM DMG

2.0 mM DMG

3.0 mM DMG

0 0.25 2.5 25 250 500

0.12 ( 0.09 0.10 ( 0.01 0.053 ( 0.002 0.025 ( 0.003 0.0041 ( 0.0010 0.0010 ( 0.0006

0.12 ( 0.020 0.097 ( 0.003 0.074 ( 0.008 0.055 ( 0.040 0.029 ( 0.008 0.030 ( 0.005

0.12 ( 0.02 0.10 ( 0.01 0.11 ( 0.04 0.12 ( 0.04 0.079 ( 0.004 0.067 ( 0.02

0.12 ( 0.08 0.11 ( 0.004 0.11 ( 0.03 0.12 ( 0.03 0.11 ( 0.03 0.11 ( 0.02

0.12 ( 0.03 0.12 ( 0.003 0.11 ( 0.03 0.11 ( 0.02 0.11 ( 0.02 0.11 ( 0.08

FIGURE 2. koserved dependence on DMG and citric acid concentrations for citric acid model solutions containing 400 nM Ni2+: [, no CA; 9, [CA] ) 0.25; 2, [CA] ) 2.5; 0, [CA] ) 25; O, [CA] ) 250; b, [CA] ) 500 µM.

FIGURE 3. kobserved vs 1/[CA] for citric acid model solutions containing 400 nM Ni2+ and 1 mM DMG: (s) linear regression curve; (‚‚‚) 85% confidence limits. reaction. The contribution of the adjunctive pathway became important for high [CA]/[DMG] ratios. For example, at [CA] ) 50 µM and [DMG] ) 1 mM (the highest [CA]/[DMG] ratio employed here), the rate coefficient for the disjunctive pathway kdisjunctive ) 0.0037 ( 0.0002 s-1 was comparable with kadjunctive ) 0.0059 ( 0.0003 s-1. The contribution of both pathways was equal for [CA] ≈ 30 µM. With decreasing [CA]/ [DMG] ratios, the contribution of the adjunctive pathway became less important. This means that for high concentrations of [DMG] (i.e., for the concentrations for which kobserved remained constant), the ligand-exchange reaction 1 followed the disjunctive pathway and the observed rate coefficient equalled the dissociation rate coefficient of Ni-CA complex.

FIGURE 4. kobserved dependence on DMG and fulvic acid concentrations for model solutions containing 400 nM Ni2+: 0, no FA; b, [FA]) 0.5; 2, [FA] ) 5; 9, [FA] ) 50; O, [FA] ) 200 mg/L. disj disj /k1 is equal to the conditional stability The ratio k-1 constant of the NiCA complex. Knowing k2 ) 0.12 s-1 (the overall rate coefficient for the test solution in the absence of CA), we calculated the conditional stability constant of this complex: K′NiCA ) 105.8 M-1. The conditional stability constant at pH 8.3 was also calculated using MINEQL and was found to be K′NiCA ) 106.3 M-1, which was in reasonable agreement with the above value. Fulvic Acid Model Systems. One of the main points of this work was to determine the relative concentrations of DMG and FA for which the ligand-exchange reaction followed the disjunctive pathway and the observed rate coefficients were independent of the experimental conditions and represented the average rate coefficients of the dissociation of Ni-FA complexes. We used the same experimental methodology to study the kinetics of ligand-exchange reaction between Ni-FA complexes and DMG as the competing ligand. The kinetic measurements of the model solutions were performed using various concentrations of DMG and FA, and the koverall was calculated using nonlinear regression analysis. The approximation with one component fitted the experimental data for all the Ni-FA test solutions except for the one with the lowest FA and DMG concentrations. For this case, the exponential growth with two components fitted the experimental data better. Hence, these data were not included in Figure 4. The results are presented in Figure 4, which shows that the overall rate coefficient strongly depended on the FA and the DMG concentrations in the test solutions when [DMG] e 3 mM. For [FA] > 50 mgL-1, the kobserved depended on [DMG] even if [DMG] > 3 mM. To find out experimental conditions for which the contributions of the adjunctive and the disjunctive pathway were significant, the reaction was

VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1087

FIGURE 5. kobserved vs 1/[FA] for fulvic acid model solutions containing 400 nM Ni2+ and 1 mM DMG: (s) linear regression curve; (‚‚‚) 95% confidence limits. studied under the experimental conditions in which the [DMG] was less than 3 mM. Figure 5 presents the kobserved versus 1/[FA] for [DMG] ) 1 mM; the intercept represents the rate coefficient of the adjunctive pathway, and the slope gives the rate coefficient of the disjunctive pathway. The rate coefficient for the ligandexchange reaction, kobserved, under the experimental conditions prescribed earlier can be obtained as follows:

1 kobserved ) 0.19 + 0.0096 [FA]

(13)

where [FA] is in milligrams per liter. From this equation, the contributions of both pathways were calculated to be equal for FA concentration ≈ 20 mg/L or [DMG]/[FA] ratio ≈ 0.05 mmol DMG/mg of FA. With increasing [DMG]/[FA] ratios, the contribution of the adjunctive pathway became less important. For example, at [DMG]/[FA] ) 0.2 nmol of DMG/mg of FA and [DMG] ) 1 mM, kdisjunctive was four times larger than kadjunctive. If we consider that the average FA concentration in natural waters is approximately 5 mg/L (assuming that approximately 50% of DOC in natural waters is FA and the concentration of DOC is normally 5-15 mgL-1), using 3 mM DMG as the competing ligand, the kdisjunctive was found to be about 12 times higher than kadjunctive. In this case, the observed rate coefficient did not depend on [DMG] and represented the average dissociation rate coefficients of Ni-FA complexes An average conditional stability constant for the Ni-FA complex can be estimated from the slope of the eq 15. Taking the bidentate complexing capacity as 5.4 mmol/g of FA (28), the molar concentration of the FA solutions was calculated and the conditional stability constant for the Ni-FA complex was estimated to be K′Ni-FA ) 105.1 M-1. As Langford et al. (5) pointed out, ligand-exchange kinetics are always recorded under experimental conditions determined by the reagent solutions whatever the pH, concentration, or ionic strength of the sample. The species identified are those of the conditions of the sample, but the rate coefficients obtained are not rate coefficients for processes that occur under the original conditions of the sample, but for processes occurring at the standardized condition imposed by the reagent solution. However, dissociation of ML complex in the ligand-exchange reaction:

ML + X f MX + L is a fundamental process of natural waters, and if the ligandexchange reaction follows the disjunctive pathway, the theory 1088

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001

of ligand substitution processes allows for a significant degree of reliable extrapolation over concentrations and pH (5). The results obtained in the model systems demonstrate that the ligand-exchange kinetics of nickel complexes with DMG as the competing ligand can follow two possible pathways, depending on the composition of the test solution, and the relative concentrations of the initial ligand, L, and the competing ligand, DMG. If the concentration of the initial ligand, L, is high as compared to that of the competing ligand, DMG, then the contribution of the adjunctive pathway becomes important. In this case, the kinetics of the overall reaction is dominated by the formation of the intermediate binuclear complex. At low ligand concentrations of the L (or high nickel loadings), the ligand-exchange reaction proceeds by the dissociation of the initial complex, which becomes the rate-determining step, and the observed rate coefficient is equal to the rate coefficient for dissociation of the initial complex, NiL. The correspondence of the observed rate coefficients and the conditional stability constant reported for the NiCA complex supports our interpretation of the ligand-exchange reactions for NiL. Interpretation of the ligand-exchange reactions for the Ni-FA model systems should be done with caution. We have not considered the fact that in these systems there may be a distribution of rate coefficients rather than a single value and that this distribution depends on the experimental conditions such as pH, ionic strength, and temperature (13, 28, 29). The above experimental results of ligand-exchange reactions of the Ni-FA complexes with DMG as the competing ligand suggest that the reaction can follow both the adjunctive and the disjunctive pathway and that one of these pathways may predominate under certain experimental conditions. This conclusion is consistent with that of Hering and Morel (21). If the concentration of the incoming ligand, DMG, is 1 mM, the contributions of the both pathways are equal at the FA concentration ≈ 20 mg/L (≈ 10 mg/L DOC) or at the [Ni]/[FA] ratio of ≈ 20 nmol of Ni/mg of FA. When the DMG concentration is increased, or for lower FA concentrations, the ligand-exchange reaction (eq 1) follows predominantly the disjunctive pathway. In this case, the observed rate coefficients become independent of [DMG] and approach the average rate coefficient of Ni-FA dissociation. The conditional stability constant calculated from the kinetic data is 105.1 M-1, whereas Burba (3) has found the stability constants for NiHA complexes as >105.3, and Sedlak et al. (27) has found it to be 103.1-104.2 M-1. Since conditional stability constants should be compared only under identical experimental conditions (which is obviously not the present case), the agreement between these values seems reasonable. Langford et al. (5, 13) and Cabaniss (7) have accepted (although not explicitly stated) that the ligand-exchange reactions between the Ni-FA complexes and PAR as the competing ligand follow the disjunctive pathway, and this would be expected for higher [Ni]/[FA] ratios (more than 100 nmol of Ni2+/mg of FA) that they have used. Assessment of the contribution of the adjunctive and the disjunctive pathway in ligand-exchange reactions of nickel complexes of natural organic matter with DMG as the competing ligand is highly complex. Both pathways are likely to be important at Ni-to-humate loading occurring in natural waters. The DOC concentration for which the contributions of both pathways are equal in model systems is higher than the values reported for the typical DOC concentrations in freshwater systems (2-10 mg/L DOC) (31). However, when comparing model systems with freshwater systems, several other factors should be considered. Freshwaters are very complex systems with various chemical and physical characteristics. Several trace metals may be present and may compete for the binding sites of organic ligands (14), and the

presence of alkaline earth metals, Ca2+ and Mg2+(usually present in massive excess), may affect the kinetics of trace metal complexation in freshwaters (30, 29). Also, other organic complexants, i.e., other than humic substances (e.g., polysaccharides, proteins), may compete for binding the trace metals including the target one. All the above competitions can affect the pathway of the overall reaction. The application of AdCSV to natural waters may also cause further complications because of possible adsorption of surface-active organic substances on the surface of the hanging mercury drop electrode (33, 34). Finally, the structure of chemically isolated humic and fulvic acids are normally different from that of humic substances present in natural waters (30). Because of these limitations, the model presented in this paper should be considered only as a first approximation to assess the predominant pathway in the overall reaction for kinetic speciation of nickel using AdCSV with DMG as the competing ligand. The mechanistic model of ligand-exchange reactions presented in this paper is reasonably consistent with kinetic and equilibrium studies. The results show the crucial role of the relative concentrations of the initial ligand and the competing ligand and the nickel-to-fulvate loading in the contribution of each pathway to the overall reaction. At very low nickel-to-fulvate loadings (typical of unpolluted freshwaters), or high concentrations of FA, the adjunctive pathway predominates. At nickel-to-fulvate loadings higher than 20 nmol of Ni/mg of DOC, the disjunctive pathway predominates. In the latter case, AdCSV with DMG as the competing ligand can be advantageously used for kinetic study of nickel complexation.

Acknowledgments The following financial support is gratefully acknowledged: Nickel Producers Environmental Research Association (USA), Inco Ltd. (Canada), and Falconbridge Ltd. (Canada), for three research contracts; Natural Sciences and Engineering Research Council of Canada (NSERC) for a research grant; NSERC Metals in the Environment Research Network grant; Ontario Power Generation Inc.; and the Mining Association of Canada. J.M. is grateful to the NSERC for a postgraduate research scholarship. M.S.A.S. is grateful to the Egyptian Ministry of Higher Education for a graduate research scholarship.

Supporting Information Available The derivation of the equations for disjunctive and adjunctive pathways for ligand-exchange kinetics (1 page). This material is available free of charge via the Internet at http:// pubs.acs.org.

Literature Cited (1) Campbell, P. G. C. In Metal Speciation and Bioavailability in Aquatic Systems; Tessier, A., Turner, D. R., Eds.; IUPAC/John Wiley & Sons Ltd.: Chichester, U.K., 1995; p 45. (2) Buffle, J. Complexation Reactions in Aquatic Systems-An Analytical Approach; Ellis Horwood: London, 1990; p 118. (3) Burba, P. Fresenius J. Anal. Chem. 1994, 348, 301. (4) van den Hoop, M. A. G. T.; van Leeuwen, H. P.; Pinheiro, J. P.; Mota, A. M.; Goncalves, M. L. S Colloids Surf. A 1995, 95, 305.

(5) Langford, C. H.; Cook, R. L. Analyst 1995, 120, 591. (6) Hering, J. G.; Morel, F. M. M. Environ. Sci. Technol. 1988, 22, 1469. (7) Cabaniss, S. E. Environ. Sci. Technol. 1990, 24, 583. (8) Pihlar, B.; Valenta, P.; Nu ¨ rnberg, H. W. J. Electroanal. Chem. 1986, 214, 157. (9) Ma, F.; Janger, D.; Renman, L. Anal. Chem. 1997, 69, 1782. (10) van den Berg, C. M. G.; Nimmo M. Sci. Total Environ. 1987, 60, 185. (11) van den Berg, C. M. G. Analyst 1989, 114, 1527. (12) Achterberg, E. P.; van den Berg, C. M. G. Deep-Sea Res. 1997, 44, 693. (13) Langford, C. H.; Gutzman, D. W. Anal. Chim. Acta 1992, 256, 183. (14) Mandal, R.; Sekaly, A. L. R.; Murimboh, J.; Chakrabarti, C. L.; Back, M. H.; Gre´goire, D. C.; Schroeder, W. H. Anal. Chim. Acta 1999, 395, 309. (15) Lu, Y.; Chakrabarti, C. L.; Back, M. H.; Sekaly, A. L. R.; Gre´goire, D. C.; Schroeder, W. H. J. Anal. Atom. Spectrom. 1996, 11, 1189. (16) Lu, Y.; Chakrabarti, C. L.; Back, M. H.; Gre´goire, D. C.; Schroeder, W. H. Int. J. Environ. Anal. Chem. 1995, 60, 313. (17) Sekaly, A. L. R.; Chakrabarti, C. L.; Gre´goire, D. C. In Proceedings of the Second Biennial International Conference on Chemical Measurement and Monitoring of the Environment; Clement, R. C., Burk, B., Eds.; Ottawa, Canada, 1998; pp 611-618. (18) Lam, M. T.; Murimboh, J.; Chakrabarti, C. L. In Proceedings of the Second Biennial International Conference on Chemical Measurement and Monitoring of the Environment; Clement, R. C., Burk, B., Eds.; Ottawa, Canada, 1998; pp 589-594. (19) Lam, M. T.; Murimboh, J.; Hassan, N. M.; Chakrabarti, C. L. Anal. Chim. Acta 1999, 402, 195. (20) Gregor, J. E.; Powell, H. K. J.; Town, R. M. J. Soil Sci. 1989, 40, 661. (21) Hering, J. G.; Morel, F. M. M. Environ. Sci. Technol. 1990, 24, 242. (22) Morel, F. M. M.; Herring, J. G. Principles and Applications of Aquatic Chemistry; John Wiley & Sons: New York, 1993; p 401. (23) Herring, J. G.; Morel, F. M. M. In Aquatic Chemical KineticsReaction Rates of Processes in Natural Waters; Stumm, W., Ed.; John Wiley & Sons: New York, 1990; p 155 (24) Kalvoda, R.; Kopanica, M. Pure Appl. Chem. 1989, 61, 97. (25) Cosovic, B. In Aquatic Chemical Kinetics-Reaction Rates of Processess in Natural Waters; Stumm, W., Ed.; John Wiley & Sons: New York, 1990; p 294. (26) Gamble, D. S. Can. J. Chem. 1970, 48, 2662. (27) Gamble, D. S. Can. J. Chem. 1972, 50, 2680. (28) Gamble, D. S.; Underdown, A. W.; Langford, C. H. Anal. Chem. 1980, 52, 1901. (29) Olson, D. L.; Shuman, M. S. Geochim. Cosmochim. Acta 1985, 49, 1371. (30) Sedlak, D. L.; Phinney, J. T.; Bedsworth, W. W. Environ. Sci Technol. 1997, 31, 3010. (31) Morel, F. M. M.; Hering, J. G. Principles and Applications of Aquatic Chemistry; John Wiley & Sons: New York, 1993; pp 358-360. (32) Malcolm, R. L.; MacCarthy, P. Environ. Sci. Technol. 1988, 22, 1469. (33) Mandal, R.; Salam, M. S. A.; Murimboh J.; Hassan, N. M.; Chakrabarti, C. L.; Back, M. H.; Gre´goire, D. C. Environ. Sci. Technol. 2000, 34, 2201. (34) Plasvic, M.; Cosovic, B. Anal. Chim. Acta 1994, 284, 539.

Received for review April 24, 2000. Revised manuscript received December 18, 2000. Accepted December 18, 2000. ES001203Q

VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1089